Feeling Older and Risk of Hospitalization

Health Psychology
2016, Vol. 35, No. 6, 634 – 637
© 2016 American Psychological Association
0278-6133/16/$12.00 http://dx.doi.org/10.1037/hea0000335
BRIEF REPORT
Feeling Older and Risk of Hospitalization:
Evidence From Three Longitudinal Cohorts
Yannick Stephan
Angelina R. Sutin and Antonio Terracciano
University of Montpellier
Florida State University
Objective: Subjective age is a biopsychosocial marker of aging with a range of health-related implications. Using 3 longitudinal samples, this study examined whether subjective age predicts hospitalization
among older adults. Method: Participants were adults aged from 24 to 102 years old, drawn from the
1995–1996 and 2004 –2005 waves of the Midlife in the United States Survey (MIDUS, N ⫽ 3209), the
2008 and 2012 waves of the Health and Retirement Study (HRS, N ⫽ 3779), and the 2011 and 2013
waves of the National Health and Aging Trends Study (NHATS, N ⫽ 3418). In each sample, subjective
age and covariates were assessed at baseline and hospitalization was assessed at follow-up. Results:
Consistent across the 3 samples, participants who felt subjectively older at baseline had an increased
likelihood of hospitalization (combined effect size: 1.17, 95% CI 1.11–1.23), controlling for age, sex,
race, and education. Further adjusting for disease burden and depression reduced the magnitude of the
association between subjective age and hospitalization in the 3 samples, but it remained significant in the
MIDUS and HRS. Conclusion: This study provides consistent evidence that subjective age predicts
incident hospitalization. Subjective age assessment can help identify individuals at greater risk of
hospitalization, who may benefit from prevention and intervention efforts.
Keywords: depressive symptoms, disease burden, hospitalization, subjective age
Supplemental materials: http://dx.doi.org/10.1037/hea0000335.supp
Subjective age is a novel biopsychosocial marker of aging that
is associated with a range of health-related outcomes, independent
of chronological age. Specifically, feeling older than one’s age is
related to increased risk of depression (Choi & DiNitto, 2014),
higher disease burden (Demakakos, Gjonca, & Nazroo, 2007),
systemic inflammation (Stephan, Sutin, & Terracciano, 2015a),
and faster cognitive decline (Stephan, Sutin, Caudroit, & Terracciano, in press). In addition, an older subjective age has been
related to premature mortality (Kotter-Grühn, KleinspehnAmmerlahn, Gerstorf, & Smith, 2009). Furthermore, subjective
age is sensitive to the biological aging of critical body systems
that are involved in health and functioning (Stephan, Sutin, &
Terracciano, 2015b). In particular, markers of better muscular
and pulmonary function, and lower adiposity are reflected in a
younger subjective age (Stephan et al., 2015b). Therefore,
subjective age shows promise as an indicator of individuals at
risk for worsening health leading to hospitalization. However,
despite existing evidence for the health-related outcomes of
subjective age, no study has yet tested whether it contributes to
risk of hospitalization.
Using data from three large longitudinal cohorts that differed in
age and time of assessment, the present study tested the hypothesis
that an older subjective age is prospectively associated with an
increased risk of incident hospitalization. In addition, the study
tested the extent to which chronic conditions and depressive symptoms accounted for this association.
Hospitalization in old age is a significant life event with a range
of adverse outcomes. Hospitalization accelerates disability in activities of daily living and declines in function (Zisberg, Shadmi,
Gur-Yaish, Tonkikh, & Sinoff, 2015) as well as cognition (Wilson
et al., 2012). Such hospitalization-related declines could lead to
rehospitalization (Jencks, Williams, & Coleman, 2009), precipitate
nursing home placement (Goodwin, Howrey, Zhang, & Kuo,
2011), and culminate in higher mortality risk (Sleiman et al.,
2009). Therefore, beyond poor health, identifying the factors that
contribute to older adults’ risk of hospitalization is critical to
prevent these adverse consequences and the substantial cost of
health care service utilization. The present study examined whether
individuals’ subjective age, which refers to how old or young they feel
relative to their chronological age, could be one risk factor for hospitalization.
This article was published Online First February 11, 2016.
Yannick Stephan, EA 4556, Dynamic of Human Abilities and Health
Behaviors, University of Montpellier; Angelina R. Sutin, Department of
Behavioral Sciences and Social Medicine, College of Medicine, Florida
State University; Antonio Terracciano, Department of Geriatrics, College
of Medicine, Florida State University.
Correspondence concerning this article should be addressed to Yannick
Stephan, EA 4556, Dynamic of Human Abilities and Health Behaviors,
University of Montpellier, UFRSTAPS, 700, Avenue du Pic St Loup,
34090 Montpellier, France. E-mail: [email protected]
634
SUBJECTIVE AGE AND HOSPITALIZATION
Method
Participants
Participants were drawn from the Midlife in the United States
Survey (MIDUS), the Health and Retirement Study (HRS), and the
National Health and Aging Trends Study (NHATS). All participants provided informed consent for participation. The analytic
sample included individuals who provided complete data on the
variables of interest at baseline and follow-up. At baseline, individuals with a recent history of hospitalization, that is in the past
12 months for the MIDUS (N ⫽ 359) and the NHATS (N ⫽ 919),
and the last two years in the HRS (N ⫽ 1,281), were excluded from
the primary analyses.
The MIDUS is a national survey of noninstitutionalized Englishspeaking adults. The MIDUS I study was supported by the John D.
and Catherine T. MacArthur Foundation Research Network on
Successful Midlife Development, and the MIDUS II was supported by a grant from the National Institute on Aging (P01AG020166). The first (1995–1996, MIDUS I) and the second
(2004 –2005) waves were used in the present study. The MIDUS
survey complied with institutional review board standards of the
University of Wisconsin and of the Harvard Medical School. Data
from 3209 participants aged from 24 to 75 years (46% male, Mean
age ⫽ 47.22, SD ⫽ 12.22) were analyzed. The HRS is a nationally
representative longitudinal study of Americans aged 50 and older
and their spouses. The HRS is sponsored by the National Institute
of Aging (Grant No. NIA U01AG009740) and conducted by the
University of Michigan. Data from the 2008 and 2012 waves were
used, leaving a sample of 3779 participants with complete data
(39% male, mean age ⫽ 67.63, SD ⫽ 9.05, age range ⫽ 50 –96
years). The NHATS is a nationally representative prospective
cohort study of Medicare enrollees aged 65 years and older.
NHATS is funded by the National Institute on Aging (NIAU01AG032947) and conducted by the Johns Hopkins Bloomberg
School of Public Health. For the present study, data from the 2011
and 2013 waves were analyzed. Complete data were obtained from
3418 participants (42% male, mean age ⫽ 76.47, SD ⫽ 7.36, age
range ⫽ 65–102 years). Attrition analyses for the three samples are
reported in online supplemental material.
635
Covariates. In line with existing research on the prediction of
hospitalization (Wilson et al., 2014), age (in years), sex (coded as
1 for men and 0 for women), race (coded as 1 for white and 0 for
other), and educational level were included as demographic covariates given their association with the risk of hospitalization. Educational level was reported in years in the HRS. In MIDUS,
education was reported on a scale composed of 12 intervals corresponding to sequential educational milestones, ranging from 1
(no grade school) to 12 (doctoral level degree). In the NHATS, the
scale ranged from 1 (No schooling completed) to 9 (Master’s,
professional or doctoral degree). In addition, disease burden and
depression were included to examine whether poor physical and
psychological health could account for the association between
subjective age and hospitalization. In the three samples, the sum of
diagnosed conditions (i.e., high blood pressure, diabetes, cancer,
lung disease, heart condition, stroke, osteoporosis or arthritis) was
computed to obtain a measure of disease burden. Depressive symptoms were assessed with the Composite International Diagnostic
Interview Short Form scales (CIDI-SF; score ranging from 0 to 7;
Kessler, Andrews, Mroczek, Ustun, & Wittchen, 1998) in the
MIDUS, an 8-item version of the Centers for Epidemiologic Research Depression (CES-D; score ranging from 0 to 8; Wallace et
al., 2000) in the HRS, and the Patient Health Questionnaire-2
(PHQ-2; score ranging from 1 to 4; Kroenke, Spitzer, & Williams,
2003) in the NHATS.
Data Analysis
In each sample, logistic regressions were used to examine the
association between subjective age at baseline and risk of hospitalization at follow-up. The baseline model tested the association
between subjective age and hospitalization, controlling for age,
sex, education, and ethnicity. In the second model, depressive
symptoms and disease burden were added. The odds ratios from
the three samples were combined using a random model metaanalysis with the comprehensive meta-analysis software.
Results
Descriptive statistics for the three samples, including the rate of
hospitalization at follow-up, are presented in Table 1. Supplemen-
Measures
Subjective age. In the three samples, participants were asked
to specify, in years, how old they felt. Consistent with previous
research (Stephan et al., 2015), proportional discrepancy scores
were calculated by subtracting chronological age from felt age,
divided by chronological age. A positive value indicated an older
subjective age. Responses three standard deviations above or below the mean were considered outliers and were excluded
(Stephan et al., 2015). Based on this criteria 40 participants from
MIDUS, 64 from HRS, and 51 from NHATS were excluded.
Hospitalization. At follow-up, MIDUS participants were
asked how many times they had been hospitalized overnight in the
last 12 months. The number of hospitalization was converted to a
dichotomous variable of hospitalized versus not hospitalized. HRS
and NHATS participants were asked whether they had an overnight hospital stay (yes/no) in the last two years and 12 months,
respectively.
Table 1
Baseline Characteristics of the Samples
MIDUS
HRS
NHATS
Variable
M/%
SD
M/%
SD
M/%
SD
Age (years)
Sex (% male)
Race (% white)
Education
Subjective age
Depressive symptoms
Disease burden
Hospitalization
(% hospitalized)
47.22
46%
94%
7.24
⫺.16
.63
2.18
12.22
—
—
2.47
.15
1.73
2.22
67.63
39%
85%
12.87
⫺.17
1.13
1.76
9.05
—
—
3.00
.15
1.75
1.20
76.47
42%
77%
5.34
⫺.17
1.41
2.32
7.36
—
—
2.25
.16
.62
1.47
13%
—
24%
—
22%
—
Note. MIDUS: N ⫽ 3209; HRS: N ⫽ 3779; NHATS: N ⫽ 3418. Education, depression, disease burden, and hospitalization were assessed using
different methods in the three samples (see Method).
STEPHAN, SUTIN, AND TERRACCIANO
636
Table 2
Logistic Regression Predicting Follow-Up Hospitalization From Baseline Subjective Age
MIDUS
Predictor
Age
Sex
Race
Education
Subjective age
Depression
Disease burden
HRS
NHATS
Model 1 Odds ratio
(95% CI)
Model 2 Odds ratio
(95% CI)
Model 1 Odds ratio
(95% CI)
Model 2 Odds ratio
(95% CI)
Model 1 Odds ratio
(95% CI)
Model 2 Odds ratio
(95% CI)
1.04 (1.03–1.05)ⴱⴱⴱ
.99 (.80–1.21)
.72 (.48–1.08)
.87 (.78–.97)ⴱⴱ
1.24 (1.12–1.39)ⴱⴱⴱ
1.04 (1.03–1.05)ⴱⴱⴱ
1.05 (.85–1.31)
.74 (.49–1.12)
.89 (.80–.99)ⴱ
1.19 (1.06–1.33)ⴱⴱ
1.08 (.98–1.20)
1.09 (1.05–1.14)ⴱⴱⴱ
1.04 (1.03–1.05)ⴱⴱⴱ
1.14 (.98–1.33)
1.08 (.87–1.35)
1.00 (.97–.1.02)
1.17 (1.08–1.26)ⴱⴱⴱ
1.04 (1.03–1.05)ⴱⴱⴱ
1.15 (.98–1.35)
1.16 (.93–1.46)
1.01 (.98–1.04)
1.11 (1.02–1.20)ⴱ
1.09 (1.01–1.18)ⴱ
1.28 (1.20–1.37)ⴱⴱⴱ
1.03 (1.01–1.04)ⴱⴱⴱ
1.09 (.92–1.29)
.84 (.70–1.03)
.87 (.80–.95)ⴱⴱ
1.12 (1.03–1.21)ⴱ
1.02 (1.01–1.04)ⴱⴱⴱ
1.18 (.99–1.40)
.88 (.72–1.07)
.91 (.84–1.00)ⴱ
1.05 (.96–1.14)
1.15 (1.06–1.24)ⴱⴱ
1.21 (1.14–1.28)ⴱⴱⴱ
Note. MIDUS: N ⫽ 3209; HRS: N ⫽ 3779; NHATS: N ⫽ 3418.
ⴱ
p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⬍ .001.
tal analysis revealed that prior experience of hospitalization was
associated with an older subjective age at baseline in the MIDUS
and the HRS but not in the NHATS (supplemental Table 1). In the
three samples, participants who felt subjectively older at baseline
had an increased likelihood of hospitalization, controlling for
demographic covariates (see Table 2). Overall, the results suggested that for every standard deviation increase in an older
subjective age (e.g., an increased tendency to feel older than one’s
age) at baseline, the risk of future hospitalization increased by 10%
(NHATS) to almost 25% (MIDUS). A random effect meta-analysis
of the three samples produced a combined effect size of 1.17 (95%
CI: 1.11–1.23), with little variation across samples (Q ⫽ 2.30, p ⫽
.32, I2 ⫽ 13). The analyses were repeated without excluding
individuals with recent hospitalizations at baseline, and revealed
that the association between subjective age and future hospitalization persisted when controlling for prior hospitalization (supplemental Table 2). Adjusting for disease burden and depression
reduced the magnitude of the association between subjective age
and hospitalization in the three samples (see Table 2), but it
remained significant in the MIDUS, the HRS, and the three
samples combined (1.11, 95% CI: 1.04 –1.18; Q ⫽ 3.01, p ⫽
.22, I2 ⫽ 33).
Discussion
Using three large longitudinal cohorts of older adults, the present study tested whether subjective age is associated with hospitalization. As expected, the results revealed that an older subjective
age predicted higher risk of incident hospitalization, independent
of chronological age and other demographic factors. This association was consistent across the three samples and different time
periods.
This study provides novel evidence that subjective age is a risk
factor for hospitalization. The health and psychological correlates
of subjective age may explain this association. Indeed, an older
subjective age is predictive of worse physical and mental health
(Choi & DiNitto, 2014; Demakakos et al., 2007), which may
necessitate health service use. Additional analysis confirmed this
assumption and revealed that disease burden and depressive symptoms accounted for part of the association between subjective age
and hospitalization. Disease burden and depressive symptoms had
a noticeable impact especially in the NHATS sample, which was
the oldest of the three samples. There could also be other biolog-
ical pathways, in addition to diagnosed diseases, that explain this
association. The higher inflammation associated with feeling older
(Stephan et al., 2015a) increases vulnerability to acute conditions
that require hospitalization. Subjective age is also sensitive to
nonpathophysiological processes indicative of muscular, pulmonary, and metabolic functions (Stephan et al., 2015b) that are likely
to convert to illness and health service use over time. Cognitive
processes are also likely to operate. An older subjective age, for
example, is associated with lower cognition and steeper cognitive
decline (Stephan et al., 2015), which have recently been related to
an increased rate of hospitalization (Wilson et al., 2014). Finally,
individuals who feel older than their age are more likely to be
sedentary (Stephan et al., 2015), which may amplify the risk of
developing or worsening chronic conditions leading to hospital
stay.
Taken as a whole, this study suggests that subjective age, with
demographic, cognitive, behavioral, and health-related factors, could
be a valuable tool to help identify individuals at risk of future
hospitalization. Individuals with an older subjective age may benefit from standard interventions, for example through physical
activity and exercise programs, which may reduce their risk of
depression and chronic disease, and ultimately their hospitalization
risk. In addition, such programs may also directly target subjective
age. Indeed, it is likely that exercise and physical activity may
promote a younger subjective age, by fostering its determinants
such as respiratory and muscular functions, self-rated health, and
positive affect (Stephan et al., 2015b). Successful activity programs may also challenge negative aging stereotypes and promote
positive attitudes toward aging and youthful self-perceptions. Future research could test whether programs that promote a more
active lifestyle have an impact on subjective age and ultimately on
hospitalization risk.
The current study had several limitations that should be considered when interpreting the results. In the three samples, hospitalization was self-reported. Hospitalization is a significant event that
should be recalled with some accuracy, but self-report biases are
still possible. Future research is needed to replicate the present
findings using Medicare claims among those aged 65 years and
older. In addition, this study focused only on the predictive role of
subjective age on hospitalization. Supplemental analyses did suggest that hospitalization may lead individuals to feel older; further
research is needed to identify the reciprocal relations between
SUBJECTIVE AGE AND HOSPITALIZATION
subjective age and hospitalization over time. Despite these limitations, this study provides new evidence that subjective age, a
biopsychosocial marker of aging, is a consistent predictor of hospitalization across three large national samples of middle age and
older adults.
References
Choi, N. G., & DiNitto, D. M. (2014). Felt age and cognitive-affective
depressive symptoms in late life. Aging & Mental Health, 18, 833– 837.
http://dx.doi.org/10.1080/13607863.2014.886669
Demakakos, P., Gjonca, E., & Nazroo, J. (2007). Age identity, age perceptions, and health: Evidence from the English Longitudinal Study of
Ageing. Annals of the New York Academy of Sciences, 1114, 279 –287.
http://dx.doi.org/10.1196/annals.1396.021
Goodwin, J. S., Howrey, B., Zhang, D. D., & Kuo, Y. F. (2011). Risk of
continued institutionalization after hospitalization in older adults. The
Journals of Gerontology: Series A: Biological Sciences and Medical
Sciences, 66, 1321–1327. http://dx.doi.org/10.1093/gerona/glr171
Jencks, S. F., Williams, M. V., & Coleman, E. A. (2009). Rehospitalizations among patients in the Medicare fee-for-service program. The New
England Journal of Medicine, 360, 1418 –1428. http://dx.doi.org/10
.1056/NEJMsa0803563
Kessler, R. C., Andrews, G., Mroczek, D., Ustun, T. B., & Wittchen, H.-U.
(1998). The World Health Organization Composite International Diagnostic Interview Short Form (CIDI-SF). International Journal of Methods in Psychiatric Research, 7, 171–185. http://dx.doi.org/10.1002/
mpr.47
Kotter-Grühn, D., Kleinspehn-Ammerlahn, A., Gerstorf, D., & Smith, J.
(2009). Self-perceptions of aging predict mortality and change with
approaching death: 16-year longitudinal results from the Berlin Aging
Study. Psychology and Aging, 24, 654 – 667. http://dx.doi.org/10.1037/
a0016510
Kroenke, K., Spitzer, R. L., & Williams, J. B. (2003). The Patient Health
Questionnaire-2: Validity of a two-item depression screener. Medical
Care, 41, 1284 –1292. http://dx.doi.org/10.1097/01.MLR.0000093487
.78664.3C
637
Sleiman, I., Rozzini, R., Barbisoni, P., Morandi, A., Ricci, A., Giordano,
A., & Trabucchi, M. (2009). Functional trajectories during hospitalization: A prognostic sign for elderly patients. The Journals of Gerontology: Series A: Biological Sciences and Medical Sciences, 64, 659 – 663.
http://dx.doi.org/10.1093/gerona/glp015
Stephan, Y., Sutin, A. R., Caudroit, J., & Terracciano, A. (2015). Subjective age and changes in memory in older adults. The Journals of
Gerontology. Series B, Psychological Sciences and Social Sciences.
Advance online publication. http://dx.doi.org/10.1093/geronb/gbv010
Stephan, Y., Sutin, A. R., & Terracciano, A. (2015a). Younger subjective
age is associated with lower C-reactive protein among older adults.
Brain, Behavior, and Immunity, 43, 33–36. http://dx.doi.org/10.1016/j
.bbi.2014.07.019
Stephan, Y., Sutin, A. R., & Terracciano, A. (2015b). How old do you feel?
The role of age discrimination and biological aging in subjective age.
PLoS ONE, 10, e0119293. http://dx.doi.org/10.1371/journal.pone
.0119293
Wallace, R., Herzog, A. R., Ofstedal, M. B., Steffick, D., Fonda, S., &
Langa, K. (2000). Documentation of affective functioning measures in
the health and retirement study. Ann Arbor, MI: Survey Research
Center, University of Michigan.
Wilson, R. S., Hebert, L. E., Scherr, P. A., Dong, X., Leurgens, S. E., &
Evans, D. A. (2012). Cognitive decline after hospitalization in a community population of older persons. Neurology, 78, 950 –956. http://dx
.doi.org/10.1212/WNL.0b013e31824d5894
Wilson, R. S., Rajan, K. B., Barnes, L. L., Hebert, L. E., Mendes de Leon,
C. F., & Evans, D. A. (2014). Cognitive aging and rate of hospitalization
in an urban population of older people. The Journals of Gerontology:
Series A: Biological Sciences and Medical Sciences, 69, 447– 454.
http://dx.doi.org/10.1093/gerona/glt145
Zisberg, A., Shadmi, E., Gur-Yaish, N., Tonkikh, O., & Sinoff, G. (2015).
Hospital-associated functional decline: The role of hospitalization processes beyond individual risk factors. Journal of the American Geriatrics Society, 63, 55– 62. http://dx.doi.org/10.1111/jgs.13193
Received June 4, 2015
Revision received November 2, 2015
Accepted November 17, 2015 䡲